The use of drug-eluting stents (DESs) in primary percutaneous coronary intervention (PPCI) has shown early benefit over bare-metal stents (BMSs) in decreasing adverse cardiac events. However, there are concerns regarding the increased risk of late and very late stent thrombosis (ST) after DES use. With the paucity of ST events individual trials may have been underpowered to detect significant differences. We sought to perform a meta-analysis to evaluate the available literature examining the outcomes of DESs and BMSs in PPCI after ≥3 years of follow-up. We analyzed 8 randomized clinical trials (RCTs) and 5 observational studies comparing DESs to BMSs in PPCI. Clinical end-point data were analyzed for RCTs and observational studies separately using random-effect models. RCTs included 5,797 patients in whom first-generation DESs (sirolimus- or paclitaxel-eluting stents) were compared to BMS control arms. Patients receiving DESs had a significantly lower risk of target lesion revascularization (odds ratio [OR] 0.48, confidence interval [CI] 0.37 to 0.61), target vessel revascularization (OR 0.53, CI 0.42 to 0.66), and accordingly major adverse cardiac events (OR 0.69; CI 0.56 to 0.84). Incidence of ST was not different between groups (OR 1.02, CI 0.76 to 1.37). There was no significant difference in mortality (OR 0.88, CI 0.68 to 1.12) or recurrent myocardial infarction (OR 0.97; CI 0.61 to 1.54). Among observational studies (n = 4,650) fewer studies reported on target lesion revascularization and target vessel revascularization, but the trend remained in favor of DESs. A small but statistically significant increase in ST was noted with DES use (OR 1.62, CI 1.18 to 2.21) at ≥3 years of follow up, without evidence of recurrent myocardial infarction. Those receiving DESs had a significantly lower mortality compared to those receiving BMSs (OR, 0.65, 95% CI 0.53 to 0.80, p <0.001). In conclusion, this meta-analysis of RCTs examining the long-term outcomes of first-generation DESs versus BMSs in PPCI, DES use resulted in decreased repeat revascularization with no increase in ST, mortality, or recurrent myocardial infarction.
The use of bare-metal stents (BMSs) in primary percutaneous coronary intervention (PPCI) in ST-segment elevation myocardial infarction has been proved to decrease rates of repeat revascularization versus balloon angioplasty alone. Over the previous 5 years multiple randomized trials and large patient registries have examined the influence of first-generation drug-eluting stent (DES) use in PPCI. Individual studies and early meta-analyses have reported an overall favorable impact of DESs compared to BMSs on target lesion revascularization and/or target vessel revascularization with no evidence of increased stent thrombosis (ST). Most reports have focused on early outcomes (≤12 months) but recently data have become available on longer-term follow-up (beyond the first year). Some of these reports have questioned the long-term efficacy of DESs in decreasing target lesion revascularization and target vessel revascularization, whereas others have raised concerns regarding an increased risk of very late ST. The paucity of events (particularly ST) in each report did not allow robust statistical analysis to clearly discern the impact of DES use in PPCI. To overcome this limitation, we performed a systematic review and meta-analysis of published reports of randomized controlled trials (RCTs) and observational studies comparing DES to BMS use in PPCI and providing clinical follow-up for ≥3 years.
Methods
Our analysis sought to answer the following questions: (1) Does DES use in PPCI decrease long-term (≥3 years) clinical adverse events compared to BMSs? (2) Is the use of DESs safe in the setting of PPCI? We report this protocol-driven systematic review and meta-analysis according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA).
Two reviewers (E.W. and A.A.L.) independently judged the eligibility of all studies. Eligible studies included RCTs and observational studies examining the use of DESs versus BMSs during PPCI that provided long-term follow-up (≥3 years). We excluded studies that did not include a control BMS arm and those that reported <3 years of follow-up.
We searched MEDLINE (January 1980 to February 2011), the Cochrane databases (January 2011), EMBASE (January 1980 to February 2011), CINAHL (January 1982 to February 2011), the United States Food and Drug Administration Web site ( http://www.fda.gov ), and BIOSIS Previews (January 1980 to February 2011) using the following database-appropriate Medical Subject Heading terms: primary percutaneous coronary intervention, balloon angioplasty, stenting, ST-segment elevation myocardial infarction, and clinical outcomes. We sought additional studies by reviewing the reference lists of eligible studies, relevant review articles, and published abstracts of major international annual meetings. Two reviewers (E.W. and A.A.L.) working in duplicate and independently used a standardized form to abstract the data from each study. K.M.Z. solved disagreements that could not be solved by consensus. For each outcome, absolute event numbers were included and results are expressed as a ratio of total participants with complete follow-up. The longest follow-up data available were used for each study.
We used the criteria of Jüni et al to ascertain methodologic quality and ascertain the potential for bias of included randomized trials and a modified Newcastle-Ottawa scale to assess the quality of registry studies (details included in supplemental material ).
Given the inherent difference in study design we performed separate meta-analyses for the RCTs and the observational studies. Prespecified outcomes of our analyses were death, myocardial infarction, ST, target vessel revascularization, target lesion revascularization, and major adverse cardiac events. Because of the variability of the definition of the composite of major adverse cardiac events, we included only studies that specifically reported the outcome and used a traditional definition of its components. For mortality some studies used all-cause mortality, whereas others used cardiac mortality. For the outcome describing need for revascularization some studies reported target vessel revascularization, whereas most used target lesion revascularization. Although there is overlap we report target lesion revascularization and target vessel revascularization outcomes to avoid excluding any study eligible for inclusion. ST was reported in accordance with Academic Research Consortium criteria. Some studies reported ST rates as definite, whereas most used definite or probable. One study reported the total of definite, probable, or possible.
Given the observed heterogeneity in the studies’ methods and types of BMS and DES used, we conducted random-effects meta-analyses to obtain estimated odds ratios (ORs) for the prespecified main clinical outcomes comparing DES- to BMS-treated patients and their associated 95% confidence intervals (CIs). The estimated OR from separate studies was combined according to the DerSemonian–Laird method.
We also estimated absolute risk decrease and numbers needed to treat to assess the clinical relevance of the results. We estimated the proportion of between-study inconsistency owing to true differences between studies (rather than differences owing to chance) using the I 2 statistic, with values of 25%, 50%, and 75% considered low, moderate, and high, respectively. Funnel plots graphically explored publication bias. RevMan 4.2 (Review Manager, Nordic Cochrane Centre, Cochrane Collaboration, Copenhagen, Denmark) was used for analyses including the estimation of absolute risk decrease.
Results
As shown in Figure 1 , 13 studies (8 RCTs and 5 observational) were found eligible for inclusion. Inter-reviewer agreement on study eligibility was 100%. Included trials and main characteristics of the 10,459 patients are presented in Table 1 . RCTs and observational studies included 5,809 patients (3,706 in the DES and 2,103 in the BMS arms) and 4,650 patients (2,613 in the DES and 2,037 in the BMS arms), respectively. In total DES-treated patients, paclitaxel-eluting stents were used in 3,880, sirolimus-eluting stents in 1,405, and a smaller number of patients received second-generation DESs.
| Study | Year | Total Number of Patients | Type of DES | Follow-Up (years) ⁎ | Age (years), mean ± SD | Women (%) | Diabetics (%) |
|---|---|---|---|---|---|---|---|
| Randomized trials | |||||||
| DEDICATION | 2010 | 626 | SES, PES | 3 | 62 DES | 27 DES | 9 DES |
| 63 BMS | 26 BMS | 11 BMS | |||||
| HORIZONS-AMI | 2010 | 3,006 | PES | 3 | 60 DES | 23 DES | 16 DES |
| 59 BMS | 24 BMS | 15 BMS | |||||
| MISSION! | 2010 | 310 | SES | 3 | 59 ± 11 DES | 25 DES | 13 DES |
| 59 ± 12 BMS | 19 BMS | 7 BMS | |||||
| PASEO | 2009 | 270 | SES, PES | 3.4 | 63 ± 15 DES | 31 DES | 23 DES |
| 62 ± 17 BMS | 29 BMS | 26 BMS | |||||
| PASSION | 2011 | 619 | PES | 5 | 61 ± 12 DES | 26 DES | 10 DES |
| 61 ± 13 BMS | 22 BMS | 12 BMS | |||||
| SESAMI | 2010 | 313 | SES | 3 | 63 DES | 20 DES | 18 DES |
| 62 BMS | 20 BMS | 24 BMS | |||||
| STRATEGY | 2009 | 175 | SES | 5 | 62 DES | 23 DES | 17 DES |
| 63 BMS | 31 BMS | 12 BMS | |||||
| TYPHOON | 2011 | 478 | SES | 4 | 58 ± 12 DES | 21 DES | 16 DES |
| 61 ± 12 BMS | 22 BMS | 17 BMS | |||||
| Observational studies | |||||||
| BASKET | 2009 | 210 | SES, PES | 3 | 62 ± 13 | 20 | 16 SES |
| 21 BMS | |||||||
| Brodie et al | 2011 | 1,463 | SES, PES | NR | NR | 28 DES | 20 DES |
| 32 BMS | 14 BMS † | ||||||
| Ishikawa et al | 2010 | 555 | SES | 3.6 DES | 67 ± 12 DES | 29 DES | 41 DES |
| 5.0 BMS † | 66 ± 12 BMS | 21 BMS | 38 BMS | ||||
| Kukreja et al | 2008 | 1,738 | SES, PES | 4.2 SES | 59 ± 12 SES | 25 SES | 12 SES |
| 2.4 PES | 60 ± 12 PES | 22 PES | 10 PES | ||||
| 5.8 BMS † | 58 ± 12 BMS † | 19 BMS | 10 BMS | ||||
| Park et al | 2010 | 684 | SES, PES | 2.1 DES | 62 ± 13 DES | 27 DES | 28 DES |
| 2.9 BMS | 62 ± 13 BMS | 22 BMS | 29 BMS |
⁎ Many trials listed event rates at a defined interval and mean ± SD was not available.
In the meta-analysis of the 8 RCTs, DES use was associated with significantly lower odds of target lesion revascularization (OR 0.48, 95% CI 0.37 to 0.61, p <0.001) and major adverse cardiac events (OR 0.67; 0.56 to 0.79, p <0.001) at ≥3 years of follow-up ( Figures 2 and 3 ) . Therefore, DES use resulted in an absolute decrease in target lesion revascularization (absolute risk decrease 7%, 95% CI 4 to 10, number needed to treat 14) and major adverse cardiac events (absolute risk decrease 7%, CI 4 to 10, number needed to treat 14). When target vessel revascularization was reported, DES use was associated with a significant decrease in that outcome (OR 0.53, 95% CI 0.42 to 0.66, p <0.001) with a significant absolute risk decrease of 8% (CI 5 to 11, number needed to treat 13). There was no statistically significant difference in mortality after ≥3 years of follow-up between DES- and BMS-treated patients (OR 0.88, 95% CI 0.68 to 1.12, p = 0.30; Figure 4 ) . Similarly, there was no significant difference between groups in incidence of ST (OR 1.02, 95% CI 0.76 to 1.37, p = 0.90; Figure 5 ) or recurrent myocardial infarction (OR 0.97, 95% CI 0.61 to 1.54; Figure 6 ) .
Given the large number of subjects and events in Harmonizing Outcomes with Revascularization and Stents in Acute Myocardial Infarction (HORIZONS-AMI), that study alone represented about 1/3 of the weight of the meta-analysis of RCTs. To explore whether overall results were primarily driven by this study, we performed sensitivity analyses for all primary outcomes excluding data from HORIZONS-AMI. Overall, there were no significant differences in the outcomes after excluding the HORIZONS-AMI data, indicating that its findings were consistent with findings of other RCTs. A similar sensitivity analysis excluding data from the Trial to Assess the Use of the Cypher Sirolimus-Eluting Coronary Stent in Acute Myocardial Infarction Treated with Balloon Angioplasty (TYPHOON), which had a 29% loss of follow-up, was performed. That analysis also did not yield different results from the overall results detailed earlier.
When the RCT data were stratified by DES type, we noted trends toward larger decreases in target lesion revascularization, target vessel revascularization, and, hence, major adverse cardiac events with sirolimus-eluting stents compared to paclitaxel-eluting stents ( Table 2 ). Similar trends were also noted in recurrent myocardial infarction and ST outcomes but not mortality.
| Outcome | OR (95% CI) | OR (95% CI) | p Value for Interaction |
|---|---|---|---|
| DES vs BMS RCTs | DES vs BMS OSs | ||
| Mortality | 0.88 (0.68–1.12) | 0.65 (0.53–0.80) | 0.07 |
| Target vessel revascularization | 0.53 (0.42–0.66) | 0.60 (0.31–1.17) | 0.73 |
| Target lesion revascularization | 0.48 (0.37–0.61) | 0.52 (0.18–1.48) | 0.90 |
| Myocardial infarction | 0.97 (0.61–1.54) | 1.11 (0.63–1.95) | 0.71 |
| Stent thrombosis | 1.02 (0.76–1.37) | 1.62 (1.18–2.21) | 0.04 |
| SES vs BMS RCTs | PES vs BMS RCTs | ||
| Mortality | 0.74 (0.49–1.11) | 0.80 (0.61–1.06) | 0.76 |
| Target vessel revascularization | 0.46 (0.34–0.64) | 0.66 (0.53–0.84) | 0.07 |
| Target lesion revascularization | 0.41 (0.29–0.59) | 0.57 (0.40–0.81) | 0.20 |
| Myocardial infarction | 0.79 (0.51–1.25) | 1.53 (0.46–5.08) | 0.31 |
| Stent thrombosis | 0.98 (0.56–1.74) | 1.12 (0.78–1.62) | 0.70 |
| Major adverse cardiac events | 0.53 (0.42–0.68) | 0.70 (0.58–0.86) | 0.08 |
| SES vs BMS OSs | PES vs BMS OSs | ||
| Mortality | 0.55 (0.37–0.80) | 0.68 (0.48–0.96) | 0.42 |
| Target vessel revascularization | 0.86 (0.45–1.63) | 0.85 (0.57–1.26) | 0.98 |
| Target lesion revascularization | 0.95 (0.46–1.99) | 0.86 (0.54–1.37) | 0.82 |
| Myocardial infarction | 0.77 (0.45–1.31) | 1.28 (0.78–2.10) | 0.17 |
| Stent thrombosis | 1.23 (0.66–2.28) | 1.49 (0.97–2.31) | 0.61 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
